From: Stuart Langridge (aquarius-lists@kryogenix.org)
Date: Mon Sep 09 2002 - 04:21:58 BST
Simon Willison spoo'd forth:
>
> Hi all,
>
> After only a few days of discussion (and today has been pretty frenetic, my
> apologies to anyone who's having trouble keeping up ;) )
You're not kidding, dude. I take two days off to actually get paid for
work and come back to loads of scratch-your-head important posts that I
have to think about at 4am. This is not helping my blood pressure ;)
> we appear to have
> hit that all-important question: How complex is too complex?
[snip]
> Here's my "case for complexity".
I shall present my case for simplicity.
There are four different sorts of bloggers in the world.
1. People who have written their own blogging system.
2. People using a pre-packaged server-side blogging system such as b2,
Movable Type, or Greymatter.
3. People using a remote interactive service that edits your static
website, such as blogger.
4. People not using a blogging system at all, but just manually editing
static pages.
Simon's case for complexity is, ironically (although arguably) based in
the assertion that XML-RPC is too complex! A RESTish interface based on
GET, or any one of a half-dozen other methods, might be simpler for
developers to implement than XML-RPC based pings. I do not believe this
to be the case, if we look at the above classes of people.
1. Certainly anyone capable of implementing their own blogging system
can also implement a method on XML-RPC for it -- I submit Simon, Hixie,
and Aquarion, all of whom managed it in jolly short order.
2. Those using pre-packaged blogs don't *have* to -- htat's why they're
using a pre-packaged blog. All it needs is for one person to be able to
implement it and then everyone else using that blogging system gets it
for free.
3. True, people using Blogger can't necessarily either make pings or
host pingbacks. If it becomes widely adopted, though, I can see Blogger
users either using a form-based interface in some central location
(which could be done based around something like Hixie's proxy, and
hosted on Simon's pingback.net), or pingback client facility being
wound into Blogger itself. For hosting pingbacks (the server side of
the Pingback spec), there would certainly be no problem with a
centralised service, in much the same way as YACCS provides a
commenting service to people with no server-side scripting facilities
on their blogs.
4. In my experience, people writing like this are inclined to not
implement much of what might be considered "standard" blog things like
permalinks, but they could certainly take the approaches described
under (3).
I think the distinction, as I've said elsewhere, is that these other
methods of pingback should be being implemented as layers on top of
Pingback rather than as an integral part of the spec. I can see users
who are writing Pingback code but don't want to use XML-RPC doing a GET
on http://pingback.net/pingback?from=http://foo&to=http://barr
example. Or going to a form on that server which then calls that very
page.
sil
--
Something must be Done
This is Something
Therefore, This must be Done
-- The Thatcherite Syllogism
Message sent over the Blogite mailing list.
Archives: http://www.aquarionics.com/misc/archives/blogite/
Instructions: http://www.aquarionics.com/misc/blogite/
This archive was generated by hypermail 2.1.5 : Tue Sep 10 2002 - 17:05:00 BST